Road defect detection

The project will further develop and optimise the road defect detection algorithm developed within ARNL. The algorithm uses a self-learning neural network, LiDAR and camera fusion to determine road surface deviations in front of vehicles that are important for vehicle speed. The resulting system warns the driver if a speed reduction is required, or generates an intervention signal to the vehicle control system in the case of a self-driving vehicle. The off-the-shelf system uses low-cost and compact camera and LiDAR devices, and the algorithm can process real-time synchronised LiDAR camera images on embedded low-power systems.